artificial irfan

AI | ML | Data with a little bit of infra and dev or vice versa.

[artificial irfan @irfanprogress]

Hi I'm Irfan, I'm your option.

[artificial irfan @irfanprogress]

@farisawalla nak collab

[Julien Chaumond @julien_c]

open source ai must win

[Yoshik @AskYoshik]

- 1 hr of learning Kubernetes
- 1 hr of grinding AWS certs
- 1 hr of building AI projects

do this straight for a year and you'll still be replaceable

[artificial irfan @irfanprogress]

@iqbalmaestro8 takkan survey je, beli terus. investment alaf baru

[artificial irfan @irfanprogress]

@ccodddemxr just do it!

[Abhishek Yadav @yabhishekhd]

Google is forcing Android apps to use less RAM. 📱

Starting February 2027, Android apps will have to follow stricter memory usage rules.

Benefits:
→ Better multitasking
→ Fewer apps reloading
→ Smoother performance
→ Fewer slowdowns and crashes
→ Better experience on phones with less RAM

Google wants developers to make apps lighter, more efficient and better at managing RAM.

Restrictions:
- Apps must better manage temporary RAM usage
- Memory should be freed when apps move to the background
- Developers need to reduce excessive image/bitmap memory usage
- Unnecessary app code must be removed
- Apps will have to shrink their code by at least 25% using tools like R8

If an app crosses the allowed memory limits, Android can slow it down or force it to close to prevent the phone from becoming slow or crashing.

When apps use less RAM, more memory remains available for the rest of the phone.

The global memory shortage is getting worse as AI data centers consume huge amounts of memory.

So the AI-driven memory crisis is now pushing Android apps to become more efficient with the RAM smartphones already have. 🧠📱

[yash.jsx @yashmp2004]

Linux users are built different 🐧

[BFM News @NewsBFM]

Asthma cases in Malaysia have surged more than eightfold as worsening air quality from the haze continues to affect parts of the country.

The Health Ministry said 1,811 asthma cases were recorded at sentinel health facilities from Aug 23 to 29, up from 219 cases the previous week.

🧵1

[Azizulhasni Awang @AzizulAWANG]

Penjelasan dan cerita sebenar tentang misi Malaysia untuk ke LA28.

[OpenAI @OpenAI]

We’re ending our partnership with Cursor following its acquisition by SpaceX. Under our proposal, Cursor’s direct access to our models would end on November 12.

We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we’re ready to go above and beyond to support them.

https://t.co/OzuCTzUjfX

[Pi @pidotdev]

Small pi update for y'all! https://t.co/BGB27QdxPa

[Chubby♨️ @kimmonismus]

GLM-5.3 is officially live on Hugging Face.

What should run it locally:

- FP8: 10–12× H100 or 8�� H200
- 4-bit/NVFP4, roughly 390–430GB: one 512GB Mac Studio or 4× DGX Spark
- Aggressive 2-bit, roughly 230–250GB: one 256GB Mac Studio or 2× DGX Spark, with quality and context tradeoffs

The first GGUF and NVFP4 quants are already appearing.

Have fun with SOTA local AI

[Alexander Belanger @abelanger5]

My favorite trick to reduce memory overhead in a Go app is to replace all time.Time references with int64. time.Time is 24 bytes versus 8 bytes, so a 3x improvement. There's also a GC improvement since time.Time contains a *Location pointer.

We maintain a few indexes in our engine which are ordered by priority and timestamp (these are queue-related indexes). We'd like to do everything possible to ensure this index fits in memory (and only fall back to disk with enormous backlogs).

In our first implementation, we'd read these rows out of the database to construct the index and convert pgtype.Timestamptz -> time.Time. But this strategy only let us fit a few million entries in memory on a 2GB machine. After a light refactor and partially due to this fix, we're nearing hundreds of millions of slots.

This is nowhere near Cloudflare's level of optimization but was very leveraged in our case.

[Zixuan Li @ZixuanLi_]

GLM-5.3 is now available for download, local deployment, fine-tuning, and commercial use under the GLM-5.3 License.

Given the model’s advanced cybersecurity capabilities, we conducted two additional weeks of comprehensive safety evaluations before releasing the weights.

Under the current license, a security review is required only when an organization with more than $10 billion in annual revenue offers GLM-5.3 as an external model service.

https://t.co/JPyk5KNDzb

[Lydia Hallie ✨ @lydiahallie]

If you're using Claude Code for physical things (firmware, robots, sensors, home automation, etc.), I'd love to hear about your setup and what you're building!

[artificial irfan @irfanprogress]

Slowly cooking something

[artificial irfan @irfanprogress]

cloudflare models should just show which self-hosted models is free or not. But it seems mcm workers AI is free based on the request/creds. So higher models capabilities, lagi cepat la habis.

[ASNB @ASNBmy_official]

Amanah Saham Nasional Berhad mengumumkan pengagihan pendapatan untuk dana harga tetap iaitu Amanah Saham Malaysia 2 – Wawasan (ASM 2 Wawasan) bagi tahun kewangan berakhir 31 Ogos 2026.

Pengagihan pendapatan ini berjumlah RM1.31 bilion, bersamaan 5.00 sen seunit, dan merupakan pengagihan pendapatan tertinggi yang diisytiharkan oleh ASM 2 Wawasan dalam tempoh tujuh tahun yang lalu. Pengagihan ini memanfaatkan lebih 1.01 juta pemegang unit dengan pegangan sebanyak 26.3 bilion unit.

Kadar pengagihan sebanyak 5.00% ini mengatasi penanda arasnya, iaitu kadar Deposit Tetap Maybank 12 Bulan pada paras 2.01% sebanyak 299 mata asas. Sehingga 24 Ogos 2026, ASM 2 Wawasan mencatatkan pendapatan bersih terealisasi sebanyak RM1.43 bilion. Prestasi yang lebih kukuh ini mencerminkan daya tahan dana dalam mengharungi persekitaran pelaburan yang semakin tidak menentu, yang dicirikan oleh peningkatan ketegangan geopolitik, termasuk konflik yang berterusan di Timur Tengah, perubahan jangkaan kadar faedah serta ketidaktentuan pasaran yang berterusan.

Urus niaga ASM 2 Wawasan di cawangan dan ejen ASNB, termasuk melalui semua saluran dalam talian dan perbankan internet, akan ditangguhkan sementara dari 28 hingga 31 Ogos 2026 bagi memudahkan pengiraan agihan pendapatan.

Pemegang unit boleh menyambung semula urus niaga melalui portal myASNB (https://t.co/n6HylcEhmm), aplikasi mudah alih myASNB, atau di mana-mana cawangan dan ejen ASNB di seluruh negara mulai 1 September 2026.

Kenyataan akhbar: https://t.co/MFGR6lutkN

#ASNB #ASM2Wawasan

[Cloudflare @Cloudflare]

We're proud to support @openai's call for collective action on cyber defense. As AI transforms the threat landscape, securing the infrastructure we all depend on requires coordinated action across the industry.

Cloudflare has a longstanding commitment to helping organizations stay secure and resilient. We remain committed to bolstering defenses against AI-enabled attacks — open-sourcing security tools, offering training, and strengthening defenses from the start.

[Anatoli Kopadze @AnatoliKopadze]

This is the best 2 hours on graph engineering ever recorded, Andrew Ng breaking down how to build agentic knowledge graphs from scratch:

9:14 - your first working agent
33:11 - how loops actually work
1:02:46 - turning loops into working graphs
1:30:15 - agents that improve their own code
1:49:05 - a system that runs the whole thing for you

I've seen $500 courses that teach less than this video alone.

Watch it, then take it further with my step-by-step guide on graph engineering below.